Patentable/Patents/US-9602833
US-9602833

Method and apparatus for deriving temporal motion vector prediction

PublishedMarch 21, 2017
Assigneenot available in USPTO data we have
Inventorsnot available in USPTO data we have
Technical Abstract

A method and apparatus for deriving a motion vector predictor (MVP) for a motion vector (MV) of a current block of a current picture in Inter, or Merge, or Skip mode. The method selects a co-located block corresponding to a co-located picture and receives one or more reference motion vectors (MVs) of one or more co-located reference blocks associated with the co-located block. The method also determines a search set and determines a search order for the search set, if the search MV corresponding to the given reference list is not available, the search order then searches the search MV corresponding to a reference list different from the given reference list. Finally, the method determines the MVP for the current block based on the search set and the search order and provides the MVP for the current block.

Patent Claims
18 claims

Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.

Claim 1

Original Legal Text

1. A method of deriving a motion vector predictor (MVP) for a motion vector (MV) of a current block of a current picture in Inter, or Merge, or Skip mode, wherein the MV is associated with the current block and a corresponding block of a target reference picture in a given reference list, the method comprising: selecting a co-located block corresponding to a co-located picture; receiving one or more reference motion vectors (MVs) of one or more co-located reference blocks associated with the co-located block; determining a search set, wherein the search set comprises one or more search MVs associated with said one or more co-located reference blocks; determining a search order for the search set, wherein the search order starts with searching the search MV corresponding to the given reference list, if the search MV corresponding to the given reference list is not available, the search order then searches the search MV corresponding to a reference list different from the given reference list; determining the MVP for the current block based on the search set and the search order; and providing the MVP for the current block, wherein said one or more co-located reference blocks comprise an inside co-located block at center of the co-located block and an outside co-located block outside the co-located block across from a lower-right corner of the co-located block; wherein said determining the MVP searches the search MVs associated with the outside co-located block and then the inside co-located block, and the search order for each co-located reference block is from the search MV corresponding to the given reference list, to the search MV corresponding to a reference list different from the given reference list; and wherein the search order for determining the MVP is from the search MV corresponding to the given reference list, to the search MV corresponding to a reference list different from the given reference list associated with the outside co-located block; and wherein said determining the MVP further searches the one or more search MVs associated with the inside co-located block if no MVP is found or the MVP found is the same as a previous MVP derived from neighboring blocks of the current block.

Plain English Translation

A method for predicting motion vectors in video compression (Inter, Merge, Skip modes) finds a Motion Vector Predictor (MVP) for a current block by: First, find a co-located block in a co-located picture. Then, get motion vectors (MVs) from "inside" (center) and "outside" (opposite lower-right corner) co-located reference blocks. Search for MVs in a specific order: first using the same reference list as the current block, then using a different reference list if needed. The search order prioritizes the "outside" block, then the "inside" block. If no MVP is found or the found MVP is the same as a previous MVP from neighboring blocks, search the "inside" block's MVs. Finally, the MVP is provided for the current block.

Claim 2

Original Legal Text

2. The method of claim 1 , wherein the co-located picture associated with the co-located block is indicated by a flag, and information associated with the flag is incorporated in sequence level, picture level, or slice level.

Plain English Translation

The method for predicting motion vectors as described, uses a flag to indicate the co-located picture. The flag's information (presence, value, etc.) is included within the video sequence, picture, or slice level data. This allows the encoder to explicitly signal which picture is used for the co-located block.

Claim 3

Original Legal Text

3. The method of claim 1 , wherein the co-located picture associated with the co-located block is derived according to an implicit method.

Plain English Translation

The method for predicting motion vectors as described, derives the co-located picture implicitly, meaning it's not directly signaled but rather determined by a predefined rule or algorithm.

Claim 4

Original Legal Text

4. The method of claim 3 , wherein selection method for the co-located picture associated with the co-located block is determined according to the given reference list.

Plain English Translation

The method of implicitly deriving the co-located picture for motion vector prediction chooses the co-located picture based on the same reference list that's being used for the current block's motion vector. The reference list guides the selection process.

Claim 5

Original Legal Text

5. The method of claim 4 , wherein the selection method selects the co-located picture in a reference list different from the given reference list.

Plain English Translation

The method that selects a co-located picture based on a reference list chooses a co-located picture from a reference list different from the one used by the current block's motion vector.

Claim 6

Original Legal Text

6. The method of claim 1 , wherein selection method for the co-located picture associated with the co-located block is based on a temporal distance associated with said one or more reference MVs of said one or more co-located reference blocks.

Plain English Translation

The method for predicting motion vectors selects the co-located picture based on the temporal distance associated with the motion vectors of the co-located reference blocks. The temporal distance represents the time difference between the co-located picture and other reference pictures.

Claim 7

Original Legal Text

7. The method of claim 6 , wherein the selection method selects the co-located block, wherein said one or more reference MVs associated with the co-located block have a shortest temporal distance.

Plain English Translation

The method that selects a co-located picture based on temporal distance selects the co-located block where the motion vectors have the shortest temporal distance, meaning the co-located picture is temporally closest to the pictures referenced by the co-located block's motion vectors.

Claim 8

Original Legal Text

8. The method of claim 1 , wherein selection method for the co-located picture associated with the co-located block is based on a picture type of the co-located picture.

Plain English Translation

The method for predicting motion vectors selects the co-located picture based on the picture type of the co-located picture (e.g., I-frame, P-frame, B-frame). Certain picture types may be preferred or avoided for co-location.

Claim 9

Original Legal Text

9. The method of claim 1 , wherein selection method for the co-located picture associated with the co-located block is based on a layer of the co-located picture in a hierarchical group of pictures (GOP), or a picture order of the co-located picture in a GOP, or the picture order of the co-located picture in a same layer as the hierarchical GOP.

Plain English Translation

The method for predicting motion vectors selects the co-located picture based on its layer in a hierarchical Group of Pictures (GOP), its picture order within the GOP, or its picture order within the same hierarchical layer of the GOP. This leverages the structure of the GOP to select a suitable co-located picture.

Claim 10

Original Legal Text

10. The method of claim 1 , wherein the inside co-located block is located at center of the co-located block.

Plain English Translation

In the method for predicting motion vectors, the inside co-located block is located at the exact center of the co-located block.

Claim 11

Original Legal Text

11. The method of claim 1 , wherein the outside co-located block is located outside the co-located block across from a lower-right corner of the co-located block.

Plain English Translation

In the method for predicting motion vectors, the outside co-located block is located outside of the co-located block, positioned diagonally opposite the lower-right corner.

Claim 12

Original Legal Text

12. The method of claim 1 , wherein said determining the MVP searches the search MVs associated with the outside co-located block before the search MVs associated with the inside co-located block.

Plain English Translation

In the method for predicting motion vectors, the search for motion vectors (MVs) starts with the outside co-located block's MVs before considering the MVs from the inside co-located block.

Claim 13

Original Legal Text

13. The method of claim 1 , the MVP is set to zero, a default value or a pre-defined value, or removed from a candidate set of MVPs if no MVP is found or the MVP found is the same as the previous MVP after the search MVs associated with the inside co-located block are searched.

Plain English Translation

In the method for predicting motion vectors, if no suitable MVP is found after searching both inside and outside blocks, or if the found MVP duplicates a neighboring block's MVP, the MVP is set to zero, a default value, a pre-defined value, or completely removed from consideration as a candidate.

Claim 14

Original Legal Text

14. The method of claim 1 , wherein said determining the MVP searches the search MV(s) associated with a second co-located reference block if said determining the MVP finds no MVP from the search MV(s) associated with a first co-located block, or if the MVP found is the same as a previous MVP derived from neighboring blocks of the current block.

Plain English Translation

In the method for predicting motion vectors, if an MVP isn't found using the motion vectors (MVs) from a first co-located reference block, or if the found MVP duplicates a neighboring block's MVP, the search proceeds to a second co-located reference block to find a suitable MVP.

Claim 15

Original Legal Text

15. The method of claim 14 , wherein the MVP is set to zero, a default value or a pre-defined value, or removed from a candidate set of MVPs, if no MVP is found or the MVP found is the same as the previous MVP after searching the search MV(s) associated with the co-located reference blocks in the search set.

Plain English Translation

Continuing the motion vector prediction method, if no MVP is found after searching all co-located reference blocks, or if a duplicate MVP is found, then the MVP is set to zero, a default value, a pre-defined value, or removed as a candidate MVP.

Claim 16

Original Legal Text

16. The method of claim 1 , wherein the MVP is scaled according to a first temporal distance associated with the MVP selected and a second temporal distance associated with the motion vector of the current block.

Plain English Translation

In the method for predicting motion vectors, the selected MVP is scaled based on temporal distances. A first temporal distance is associated with the selected MVP, and a second temporal distance is associated with the current block's motion vector. This scaling adjusts the MVP to account for differences in temporal relationships between the frames.

Claim 17

Original Legal Text

17. An apparatus for deriving a motion vector predictor (MVP) for a motion vector (MV) of a current block of a current picture in Inter, or Merge, or Skip mode, wherein the MV is associated with the current block and a corresponding block of a target reference picture in a given reference list, the apparatus comprising at least one circuit configured for: for selecting a co-located block corresponding to a co-located picture; receiving one or more reference motion vectors (MVs) of one or more co-located reference blocks associated with the co-located block; determining a search set, wherein the search set comprises one or more search MVs associated with said one or more co-located reference blocks; determining a search order for the search set, wherein the search order starts with searching the search MV corresponding to the given reference list, if the search MV corresponding to the given reference list is not available, the search order then searches the search MV corresponding to a reference list different from the given reference list; determining the MVP for the current block based on the search set and the search order; and providing the MVP for the current block, wherein said one or more co-located reference blocks comprise an inside co-located block at center of the co-located block and an outside co-located block outside the co-located block across from a lower-right corner of the co-located block; wherein said determining the MVP searches the search MVs associated with the outside co-located block and then the inside co-located block, and the search order for each co-located reference block is from the search MV corresponding to the given reference list, to the search MV corresponding to a reference list different from the given reference list; and wherein the search order for determining the MVP is from the search MV corresponding to the given reference list, to the search MV corresponding to a reference list different from the given reference list associated with the outside co-located block; and wherein said determining the MVP further searches the one or more search MVs associated with the inside co-located block if no MVP is found or the MVP found is the same as a previous MVP derived from neighboring blocks of the current block.

Plain English Translation

An apparatus for motion vector prediction (Inter, Merge, Skip modes) contains circuits configured to: Find a co-located block in a co-located picture. Get motion vectors (MVs) from "inside" (center) and "outside" (opposite lower-right corner) co-located reference blocks. Search for MVs in a specific order: same reference list as the current block first, then a different list if needed. Prioritize the "outside" block, then the "inside" block. If no MVP is found or the found MVP is the same as a previous MVP from neighboring blocks, search the "inside" block's MVs. Finally, an MVP is provided for the current block.

Claim 18

Original Legal Text

18. The apparatus of claim 17 , wherein the co-located picture associated with the co-located block is indicated by a flag, and information associated with the flag is incorporated in a sequence level, a picture level, or a slice level.

Plain English Translation

The motion vector prediction apparatus, as described, where a flag indicates the co-located picture, and the flag information is included within the video sequence, picture, or slice level data. This allows the encoder to explicitly signal which picture is used for the co-located block.

Classification Codes (CPC)

Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.

Patent Metadata

Filing Date

March 17, 2016

Publication Date

March 21, 2017

Want to explore more patents?

Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.

Citation & reuse

Analysis on this page is generated by Patentable — an AI-powered patent intelligence platform. AI-generated summaries, explanations, FAQs, and analysis may be reused with attribution and a visible link back to the canonical URL below. Patent abstracts and claims are USPTO public domain.

Cite as: Patentable. “Method and apparatus for deriving temporal motion vector prediction” (US-9602833). https://patentable.app/patents/US-9602833

© 2026 Nomic Interactive Technology LLC. Machine-readable context available at /api/llm-context/US-9602833. See llms.txt for full attribution policy.